Geoscience Reference
In-Depth Information
Fig. 5 End-of-run bed surfaces
(tunnel soffit removed, flow
from
bottom
to
top
): (a)
Run 9 of
d
¼
0.11 mm,
u
*
/
u
*c
¼
1.27, and grain
Reynolds number
R
*
¼
u
*
d
/
v
¼
1.8 and (b) Run 11 of
d
¼
0.87 mm,
u
*
/
u
*c
¼
1.30,
and
R
*
¼
25
area of a flow (e.g. Adrian
1986
; Hesselink
1988
; Gray et al.
1991
). In order to
further examine the potential roles of fluid-sediment system instabilities or turbu-
lent fluid motions on bedform generation, PIV measurements of near-bed flow were
made for bedforms growing from plane-bed conditions.
The PIV system utilised in this study is unique in a number of respects (Schlicke
et al.
2007
). The system uses a high-speed camera to image flow fields illuminated
by a scanning-beam lightsheet, where the lightsheet is generated from a 5W CW
532 nm Nd:YVO4 laser using a galvanometer-driven mirror (computer controlled)
together with a parabolic mirror (Fig.
4b
). The system has several advantages over
rotating-polygon type scanning systems, including: that the mirror is always posi-
tioned at the focal point of the parabola, and that the lightsheet generation is
extremely versatile, with the lightsheet width (for a single parabolic mirror) and
beam scan velocities easily and independently adjustable. Additionally, the beam
scan velocity, which is typically non-linear in rotating-polygon systems due to
inherent properties of parabolic reflectors, can be constant in a galvanometer-based
system (giving a uniform intensity lightsheet) by driving the galvanometer at an
unsteady angular velocity. The system also offers additional benefits over equiva-
lent double-pulsed or twin-laser setups that rely on beam expansion by lens systems
and that typically only allow measurement at frequencies up to 50 Hz. Integrated
synchronisation options for the system permit frame-straddling techniques to be
used in order to reduce interframe times to below 1/(camera frame rate). The system
as designed is capable of resolving flow fields at frequencies of up to 200 Hz
(covering the typical scales of interest for hydraulic researchers), with recording
durations of over 8 min.
The bedform-generation experiments were made using a glass-sided tilting
recirculating (water and sediment) laboratory flume (measuring 0.44 m
0.38 m
(deep)
12 m). The tested bed was composed of a uniform quartz sediment
of median size
d
0.8 mm. The two flows tested were of respective values of
¼
u
*
/
u
*c
¼
1.8 and 1.5, and
R
*
¼
28 and 23, where the flow depth in each case was
